Use of .0 and .255 in address pools
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks? More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses? --David
On 2009-01-21, at 17:11, David Robb wrote:
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks?
More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses?
Here in AS 5645 we're avoiding handing out any v4 address whose final octet is 0 or 255, but mainly out of fear the unknown. It would be handy to see a modern study of whether stuff will still break, since being able to configure big pools instead of big piles of little ones would be a minor win. Joe
for a few years, I've put my SMTP service on a .0 address and DNS services on .255 - its been more than a decade since CIDR notation became the norm... nothing majik abt .0 or .255 - save the bounding addresses on a block delegation. --bill On Wed, Jan 21, 2009 at 05:32:03PM -0500, Joe Abley wrote:
On 2009-01-21, at 17:11, David Robb wrote:
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks?
More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses?
Here in AS 5645 we're avoiding handing out any v4 address whose final octet is 0 or 255, but mainly out of fear the unknown. It would be handy to see a modern study of whether stuff will still break, since being able to configure big pools instead of big piles of little ones would be a minor win.
Joe _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog
On 22/01/2009, at 11:44 AM, bmanning(a)vacation.karoshi.com wrote:
for a few years, I've put my SMTP service on a .0 address and DNS services on .255 - its been more than a decade since CIDR notation became the norm... nothing majik abt .0 or .255 - save the bounding addresses on a block delegation.
One host connecting to another host with a .0 or .255 address is a bit different to assigning a .0 or .255 address to a CPE with IPCP though. I haven't heard of any horror stories in a while, but, you never know. I guess if the addressing is dynamic, the user will restart their CPE and get a new address, and it fixes itself... Crude though. -- Nathan Ward
On Wed, Jan 21, 2009 at 08:32:19PM -0500, Joe Abley wrote:
On 21 Jan 2009, at 17:44, bmanning(a)vacation.karoshi.com wrote:
for a few years, I've put my SMTP service on a .0 address and DNS services on .255
Can you tell how many connection attempts or queries you're receiving from hosts that refuse to send them? :-)
none! i get no connection attempts or queries from hosts that refuse to send them. :) nxt q plz --bill
On Thu, 2009-01-22 at 11:11 +1300, David Robb wrote:
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks?
More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses?
--David
I tried this over a year ago, maybe over two. The thing which stopped me was the bank websites and their firewalling which appeared to drop any packets from these addresses. Regards, RH.
I had exactly this problem when I last used /20's etc as IP Pools. This was some time ago however, so I am hoping things have changed.
I tried this over a year ago, maybe over two. The thing which stopped me was the bank websites and their firewalling which appeared to drop any packets from these addresses.
David Robb wrote:
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks?
More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses?
I tend to use larger subnets, but exclude .0 and .255 (lots of BRAS-y type equipment can do this; or intelligent DHCP/RADIUS servers). I found there was still too much broken equipment and OS, or broken firewall admins, on the internet. I noted Vodafone handing out .0 IPs on their 3G service a few months ago, and it did indeed cause some things to break for me. aj
A year or so ago my experience was that Windows Server 2003 is not able to connect to a device with a .255 address from what was a once classified as a C network. A Linux box had no issues with the same. I would say there are still a few TCP/P stacks out there that don't fully support classless networks. Ivan Alastair Johnson wrote:
David Robb wrote:
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks?
More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses?
I tend to use larger subnets, but exclude .0 and .255 (lots of BRAS-y type equipment can do this; or intelligent DHCP/RADIUS servers). I found there was still too much broken equipment and OS, or broken firewall admins, on the internet.
I noted Vodafone handing out .0 IPs on their 3G service a few months ago, and it did indeed cause some things to break for me.
aj _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog
and you don't know the great comfort I have when the WS-2003 zombie botnet army can't find my SMTP server... :) --bill On Thu, Jan 22, 2009 at 06:12:53PM +1300, Ivan Walker wrote:
A year or so ago my experience was that Windows Server 2003 is not able to connect to a device with a .255 address from what was a once classified as a C network. A Linux box had no issues with the same. I would say there are still a few TCP/P stacks out there that don't fully support classless networks.
Ivan
Alastair Johnson wrote:
David Robb wrote:
Curiousity about what people are using as far as addressing pools these days - lots of /24s (and excluding the end addresses), or bigger blocks?
More specifically, is anyone handing out the .0 and .255 addresses, or are there still sufficient CPE/endpoints out there which cope poorly with those addresses?
I tend to use larger subnets, but exclude .0 and .255 (lots of BRAS-y type equipment can do this; or intelligent DHCP/RADIUS servers). I found there was still too much broken equipment and OS, or broken firewall admins, on the internet.
I noted Vodafone handing out .0 IPs on their 3G service a few months ago, and it did indeed cause some things to break for me.
aj _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog
_______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog
participants (8)
-
Alastair Johnson -
bmanning＠vacation.karoshi.com -
David Robb -
Ivan Walker -
Joe Abley -
Nathan Ward -
Richard Haakma -
Simon Allard